Physicochemical methods for production of enantiomerically pure compounds usually involve multi-step synthesis incorporating one or more steps which are asymmetric, and laborious purification procedures. Such methods are not only tedious, but frequently provide relatively poor yields. Alternatively enantiomerically-pure starting materials can be used, together with enantioselective reaction steps; however, such pure starting materials are available only for a very limited number of desired compounds.
In an attempt to overcome the difficulties of using traditional organic chemical methods, biological systems have been intensively investigated. Such systems show a very high degree of stereoselectivity in their reactions, and therefore microbiological, enzymatic or chemoenzymatic reactions for achieving specific reaction steps with a variety of reagents have been attempted. For example, microorganisms of a number of genera have been proposed for synthesis of optically active .alpha.-substituted derivatives of 3-hydroxypropionic acid for use as intermediates in the synthesis of compounds such as .alpha.-tocopherol, muscones and pharmaceutical, insecticidal and agricultural chemical agents (U.S. Pat. No. 4,734,367 by Hoffman-La Roche, Inc.). Most such procedures use whole-cell fermentation systems in aqueous media, or isolated enzymes with a specific desired activity. However, fermentation systems present the disadvantage that purification of the desired product can be difficult, and yields tend to be low; while the yield and convenience of the reaction can be improved by utilising immobilised cells, or cells which have been selected or genetically modified, this adds significantly to the cost of the process. The use of purified enzymes is normally prohibitively expensive, and again without the use of immobilised enzyme the yield tends to be low and purification difficult.
In recent years, intense efforts have been directed towards development of methods which are highly selective, provide a good rate of transformation, and enable easy, non-chromatographic separation and purification of the product. It would be particularly desirable if reactions could be carried out in organic solvents, since these are particularly convenient for large scale reactions and purifications.
It has been shown that dry baker's yeast is able to effect non-fermentative reduction of a-keto esters in organic solvents such as hexane or benzene, to produce the corresponding .alpha.-hydroxy esters with good yield and selectivity (Nakamura et al, 1988; Nakamura et al, 1990; Nakamura et al, 1991; Nakamura et al, 1993); reduction of .beta.-keto esters in petroleum ether, diethyl ether, toluene, carbon tetrachloride and petrol has also been demonstrated (Jayasinghe et al, 1993; Jayasinghe et al, 1994; North, 1996). Although initially it was thought that immobilisation of yeast, for example in polyurethane, was essential in order to maintain stability of cell membrane-bound coenzymes for the dehydrogenases and reductases which catalyse the reaction (Nakamura et al, 1988; Nakamura et al, 1990), it was subsequently found that the addition of a very small proportion of water to the organic system would avoid the need for immobilisation (Nakamura et al, 1991).
Ephedrine (.alpha.-[1-(methylamino)ethyl]benzene-methanol), originally isolated from plants of the genus Ephedra, occurs as the naturally-occurring isomers l-ephedrine and d-pseudoephedrine, and other pharmacologically active isomers include d-ephedrine and l-pseudoephedrine. These compounds are adrenergic sympathomimetic agents and have antihistamine activity; l-ephedrine is widely used as a bronchodilator, while d-pseudoephedrine is widely used as a decongestant. Compounds of these groups are present in a very wide range of prescription and over-the-counter pharmaceutical formulations.
The production of l-phenylacetylcarbinol, a precursor of l-ephedrine, by catalysis using whole baker's yeast cells in aqueous medium was one of the first microbial biotransformation processes to be used commercially (Neuberg and Hirsch, 1921; see also Hildebrandt and Klavehn, 1934). This reaction involves the yeast-induced condensation of benzaldehyde with acetyl-coenzyme A. The reaction has been widely investigated, and has been shown to be mediated by the enzyme pyruvate decarboxylase (Groger, Schmander and Mothes, 1966). It has also been shown that the reaction has a relatively broad specificity for the substrate, enabling a variety of substituted aromatic aldehydes to be converted to the corresponding substituted optically-active phenylacetylcarbinols (Long, James and Ward, 1989).
Although this yeast-catalysed system has been widely exploited, this has normally utilised aqueous systems, which are inconvenient for large-scale extraction and purification, which require organic solvents. Additionally, fermentation systems present the disadvantage that purification of the desired product can be difficult, and yields tend to be low; while the yield and convenience of the reaction can be improved by utilising immobilised cells, or cells which have been selected or genetically modified, this adds significantly to the cost of the process. The use of purified enzymes is normally prohibitively expensive, and again without the use of immobilised enzyme the yield tends to be low and purification difficult.
We have now surprisingly found that yeast-mediated acyloin condensation of benzaldehyde can be achieved in an organic solvent using non-fermenting yeast, and that addition of a small proportion of ethanol to the reaction mixture suppresses formation of undesired side-products. Even more surprisingly, by performing the reaction at reduced temperature, an even greater reduction of side-reactions can be achieved, without loss of catalytic activity. The effect of reduction in temperature appears to be generally applicable to both aqueous and non-aqueous systems utilising a non-fermenting yeast.
Although Ward and co-workers have carried out investigations using whole cell yeast biotransformation in two-phase organic systems with a water content of at least 10% (Nikolova and Ward, 1991; 1992a; 1992b; Ward, 1995), the yields of phenylacetylcarbinol were low, and the levels of side-products were unacceptably high.
The first description of the synthesis of l-ephedrine was contained in a patent by Hildebrandt and Klavehn (1934) and made use of the discovery by Neuburg and Hirsch (1921) that fermenting strains of Saccharomyces cerevisae in aqueous systems would convert benzaldehyde to phenylacetylcarbinol. The yield of the carbinol was typically about 18%, and significant amounts of both benzyl alcohol and benzoic acid were obtained as side-products.
In a preferred embodiment, yields of around 24% with the almost total absence of side-products were obtained using the method of the invention.